Does anybody know how to make

Observable.error(CustomException())

or even

throw CustomException()

work in Kotlin? RxJava is having trouble propagating the error due to it not being a checked exception I think:

rx.exceptions.OnErrorFailedException: Error occurred when trying to propagate error to Observer.onError

	at rx.observers.SafeSubscriber._onError(SafeSubscriber.java:192)
	at rx.observers.SafeSubscriber.onError(SafeSubscriber.java:120)
	at rx.internal.operators.OperatorSubscribeOn$1$1.onError(OperatorSubscribeOn.java:59)
	at rx.internal.operators.OnSubscribeThrow.call(OnSubscribeThrow.java:44)
	at rx.internal.operators.OnSubscribeThrow.call(OnSubscribeThrow.java:28)
	at rx.Observable.unsafeSubscribe(Observable.java:8460)
	at rx.internal.operators.OperatorSubscribeOn$1.call(OperatorSubscribeOn.java:94)
	at rx.internal.schedulers.ImmediateScheduler$InnerImmediateScheduler.schedule(ImmediateScheduler.java:57)
	at rx.internal.operators.OperatorSubscribeOn.call(OperatorSubscribeOn.java:45)
	at rx.internal.operators.OperatorSubscribeOn.call(OperatorSubscribeOn.java:30)
	at rx.internal.operators.OnSubscribeLift.call(OnSubscribeLift.java:50)
	at rx.internal.operators.OnSubscribeLift.call(OnSubscribeLift.java:30)
	at rx.Observable.subscribe(Observable.java:8553)
	at rx.Observable.subscribe(Observable.java:8520)
	at rx.Observable.subscribe(Observable.java:8375)

what do you use for dependency injection with android and kotlin? Dagger? Kodein?

Dagger 2

@darrenatherton has nothing to do with Kotlin. Your onError handler is throwing an exception while processing your error.

@jw The observable itself works outside of tests:

return if (userId != null) {
            the normal observable
        } else {
            Observable.error(NotLoggedInException())
        }

The error handling itself doesn’t do anything:

{ error -> Log.e("presenter", error.message) }

So I’m wondering if it’s my test code:

thenReturn(Observable.error(NotLoggedInException())

The only other difference is using Schedulers.immediate() in test.

One more question. I want to stay with building UI using XML instead of Anko. Should I use https://kotlinlang.org/docs/tutorials/android-plugin.html or maybe ButterKnife bindings https://github.com/JakeWharton/kotterknife ?

@jakub.dyszkiewicz Ultimately it’s down to which syntax you prefer really.

@jakub.dyszkiewicz I like to use android extensions, it looks really clean solution when you get used to it

I use the extensions

Extensions quick tip = if you wish to keep the names of some of your xml elements, you can do a rename:

import kotlinx.android.synthetic.main.content_main.recyclerView as customName

then call

customName.adapter

etc

Yeah I use it same way 🙂 That way extensions don't need to have impact on view id naming in xml files

och, that’s cool! 🙂

excellent, I will try kotlin-extensions, thanks

That being said, I don’t rename much. Instead, I just have short names in XML, as I’ve always found standard naming fairly convoluted

Anyone uses kovenant? I'm trying it right now and I've never used promises so I'm having some issues, and I can't find some reference or documentation... I can do task { "1" } and task { "2" } and it will return a promise of Tuple2, but how can I run task one and then task two? task { "1" } then { task { "2" } } seems ugly to me, moreover I can't return the second task to concat other things

but for example if you have many layouts and lot of same view ids, I get annoyed when I try to do autocomplete on R.id. 🙂 Too many suggestions 😄

The problem is when you have 2 layout with some repeated id and autoimporting synthetics takes the wrong file... But that's not a big issue compared to the advantage of "autobinding" 😁

so distinct ids solve the problem 😄

Yeah, usually I prepend the activity name if that's something related to a single activity Eg: MainActivity.kt, R.layout.activity_main, views will be R.id.main_username R.id.main_profile_img and so on (just an example)

is it a security risk to have android debugging enabled on my device? you have to have the device unlocked and confirm to trust the computer..so i don't see how it could be. there is no "trust every computer" option

Nop

Even with the dev options enabled, nobody can activate or desactivate stuff programatically. You need to have a rooted device if you want to do things like this

Hey, quick question for y’all, as a non-java non-android dev, I wonder if it’s acceptable in any way to not have such deep nesting of directories for my projects (using the deep namespacing like com/blah/appname/etc)

obviously I understand if there’s specific design reasoning behind this approach, but if there is a way to avoid it for simple things, I’d love to know

I’m really unfamiliar with the details under the hood so please forgive my ignorance

I'm not sure, so, if an experienced dev could correct me, it can be nice 😃 The "deep namespacing" allows us to develop packages without carring of if another package as the same name. For example, you may want to create your own

FileReader

package, but java has one. Then, you don't care because your package will be named com.you.

FileReader

Ahh yes, avoiding collisions, I did have that in mind. So is it really a design feature of the package system that sets the directory structure requirements?

And I guess since everything we do in android has to coexist with the entire ecosystem

@nyancat Yes, since package name is unique (in theory) imports will never conflict. There is no technical reason but you should probably keep the namespace declaration

That makes sense